Leader behavior studies have provided consistent results in the past decades. It seems that onsideration and initiating structure are two major dimensions of leader behaviors. Implicit leadership theories, however, argued that the two leader behavior dimensions are just a reflection of subordinates' implicit categorization, not the real behaviors presented by leaders. In both Mainland China and Taiwan, moral and character factors always emerged from leader behavior studies. Using a bottom-up inquiry approach, the present study developed a leader behavior inventory based on critical incident data collected in earlier studies. The inventory was administered to 750 employees. Results indicated that two factors, positive and negative leader behaviors, consistently emerged in different factor analysis procedures. Cluster analysis revealed that subjects could be properly divided into two clusters based on their responses in the leader behavior inventory. One cluster of subjects intended to evaluate their leaders positively, another one was just the opposite. The author tries to explain the results with cultural perspectives. In Chinese culture, dual­categorization (like yin/yang, good/evil) is often used in any situations. The most popular character sketch is Chun-tzu (i.e., real gentleman) and Hsiao-jen (i.e, mean person) presented in the Confucian Analects. Based on the implicit leadership theory, it is possible that subjects' responses on the leader behavior inventory just reflect their implicit categorization leader behaviors. Taking earlier related findings as evidence, the author proposes that moral evaluation is a key component in Chinese leader behavior. The cultural context is really important in indigenous leader behavior research.